Hi - I would like to know why did you need to create the server with http instead using the express app as is? as in invoke the listen method on the express and not on http.
Hi is there a way to do a performance test of a CPU extensive node app. The app is cluster app with child_process and worker thread with lot of blocking code.
Hi Vinay. If I understand your question correctly, the best would be to expose the cluster app via an endpoint and use AutoCannon as I did? Maybe I'm not understanding your requirement :)
I suggest the way how to increase a performance test quality - spin up a dedicated virtual machine (in some cloud) for your application (or spin up a dedicated stress-testing environment for your project, intended to be used only for the performance-tests), and a dedicated virtual machine (probably, more powerful) for the autocannon. By doing so your application will not be affected by the autocannon itself, and your testing condition will be 100% predictable. Well, not 100% predictable if you use Virtual machines, your results might depend on Virtualization and underlying host machine utilization. So, I suggest (and most of the time it is not cheap, unfortunately), spin up the dedicated server. One caveat to mention - your cloud provider might think that your performance tests is a DoS attack. In some situations, it is better to notify your hosting provider in advance about the traffic and purpose.
just benchmark outside of your process, VM, LAN and do the real benchmark using clients over the internet. you'll find you are practically doing the same thing with express, Fastify or even php
You are underrated, lots of good information in your videos, please don't stop making videos like this. but fastify is becoming a better alternative for express. not many good videos are out there, even matteo(founder of fastify) tweeted last week that there aren't enough videos on fastify, which is the root cause of it not growing, as most developers prefer videos. why don't you make videos on fastify?
Hi Sachin. Thank you for watching and thanks for this awesome feedback. Yes Fastify is definitely on my list...The only reason I use Express is because currently most of my client environments run Express, and I want my tutorials on performance to be factual based on production environments and stability. We are slowly moving towards Nest/Fastify and as such will be able to benchmark performance accordingly and from that, I can throw out some content, good or bad 😎.
Hi coders 👋. Hope you're all enjoying the series. Question: Which performance benchmarking tool do you use and or recommend? Cheers 😎
Looking for the next videos in 2021 brother.
These Node JS Performance Optimizations videos are a real hidden gem. You deserve more views! Keep up the amazing work.
Awesome Mark thanks for the great feedback. A video on Multithreading on its way soon 💪.
Awesome video. waiting for your next video.
Glad you enjoyed it. Have you seen my other videos on Node JS Performance Practices? 🙂
Hi - I would like to know why did you need to create the server with http instead using the express app as is? as in invoke the listen method on the express and not on http.
I went the route that most people would be familiar with.
Man how are you now famous...Thank you very much for the information. Very few do that.
Hi Sachin. Thank you for the awesome comment. It's really appreciated 🙏
can we do this inside the docker container that runs the nodejs application so that we can do this in an isolated environment?
great videos man
Awesome thanks Yonatan 🙏
Hi is there a way to do a performance test of a CPU extensive node app. The app is cluster app with child_process and worker thread with lot of blocking code.
Hi Vinay. If I understand your question correctly, the best would be to expose the cluster app via an endpoint and use AutoCannon as I did? Maybe I'm not understanding your requirement :)
whats your pc configuration? @bleeding code?
Hi Harish. I run a MacBook Air (M1, 2020) with 8GB RAM
Excellent!!
Thank you Prabhu 🙏🙂
I suggest the way how to increase a performance test quality - spin up a dedicated virtual machine (in some cloud) for your application (or spin up a dedicated stress-testing environment for your project, intended to be used only for the performance-tests), and a dedicated virtual machine (probably, more powerful) for the autocannon. By doing so your application will not be affected by the autocannon itself, and your testing condition will be 100% predictable. Well, not 100% predictable if you use Virtual machines, your results might depend on Virtualization and underlying host machine utilization. So, I suggest (and most of the time it is not cheap, unfortunately), spin up the dedicated server. One caveat to mention - your cloud provider might think that your performance tests is a DoS attack. In some situations, it is better to notify your hosting provider in advance about the traffic and purpose.
Great tip thanks Vladimir 👍
How to improve performance: Ditch express and use fastify
Yes that's definitely a good path to go and will be a video on its own in the near future. Thanks for watching 🙏
just benchmark outside of your process, VM, LAN and do the real benchmark using clients over the internet. you'll find you are practically doing the same thing with express, Fastify or even php
@@aben7810 the implementation of the thing does matter
You are underrated, lots of good information in your videos, please don't stop making videos like this. but fastify is becoming a better alternative for express. not many good videos are out there, even matteo(founder of fastify) tweeted last week that there aren't enough videos on fastify, which is the root cause of it not growing, as most developers prefer videos. why don't you make videos on fastify?
Hi Sachin. Thank you for watching and thanks for this awesome feedback. Yes Fastify is definitely on my list...The only reason I use Express is because currently most of my client environments run Express, and I want my tutorials on performance to be factual based on production environments and stability. We are slowly moving towards Nest/Fastify and as such will be able to benchmark performance accordingly and from that, I can throw out some content, good or bad 😎.
Protip: Don't log 404s unless you want a really easy way to DDoS your server/overload IO/fill your disk with logs.
Thanks for the tip 👍
Man, be more concise...
Lol. Cheers thanks :)